The Task
This takehome exercise aims to sharpen the skill of building data visualisation programmatically using appropriate tidyverse family of packages and the preparation of statistical graphics using ggplot2 and its extensions. The specific requirements can be found in the screenshot below.
You can find the links to then datasets here.
Exploration
Initialisation
Getting Packages
The code chunk below is used to install and load the required packages onto RStudio.
packages = c('sf', 'tmap', 'tidyverse',
'lubridate', 'clock', 'sftime',
'rmarkdown', 'dplyr', 'readr')
for (p in packages){
if(!require(p, character.only = T)){
install.packages(p)
}
library(p,character.only = T)
}
Getting Data
Data For The Types of Buildings and Their Max Occupancy
The code chunk below is used to load the necessary data.
buildings <- read_sf("data/takehome_exercise_5/Buildings.csv",
options = "GEOM_POSSIBLE_NAMES=location")
schools <- read_csv("data/takehome_exercise_5/Schools.csv")
pubs <- read_csv("data/takehome_exercise_5/Pubs.csv")
apartments <- read_csv("data/takehome_exercise_5/Apartments.csv")
employers <- read_csv("data/takehome_exercise_5/Employers.csv")
restaurants <- read_csv("data/takehome_exercise_5/Restaurants.csv")
checkInJournal <- read_csv("data/takehome_exercise_5/CheckInJournal.csv")
The buildings dataset represents every building in Engagement. However, it does not give enough information about what each building is. To do that, we have to combine it with the apartments, employers, pubs, restaurants and schools datasets. We do that with the code chunks below.
# Reformat datasets apartments, employers, pubs and restaurants
# Apartments
apartments <- apartments[,c("buildingId", "maxOccupancy")]
apartments$buildingType <- "Apartment"
# Employers
employers <- employers[,c("buildingId")]
employers$maxOccupancy <- ""
employers$buildingType <- "Employer"
# Pubs
pubs <- pubs[,c("buildingId", "maxOccupancy")]
pubs$buildingType <- "Pub"
# Restaurants
restaurants <- restaurants[,c("buildingId", "maxOccupancy")]
restaurants$buildingType <- "Restaurant"
# Schools
schools <- schools[,c("buildingId", "maxEnrollment")]
schools$buildingType <- "School"
names(schools)[names(schools) == 'maxEnrollment'] <- "maxOccupancy"
# Combine reformatted datasets
intermediate_buildings <- rbind(apartments, employers, pubs, restaurants, schools)
# Combine reformatted datasets
intermediate_buildings <- rbind(apartments, employers, pubs, restaurants, schools)
# Remove duplicates
intermediate_buildings_unique <- unique(intermediate_buildings)
# Now we combine the dataset intermediate_buildings_unique with the dataset buildings to give as many buildings in Engagement a category as possible
buildings_test <- merge(x = buildings, y = intermediate_buildings_unique, by = "buildingId", all.x = TRUE)
# We combine the building types
names(buildings_test)[names(buildings_test) == 'buildingType.x'] <- "Generic"
names(buildings_test)[names(buildings_test) == 'buildingType.y'] <- "Specific"
buildings_test$buildingType <- ifelse(is.na(buildings_test$Specific), buildings_test$Generic , buildings_test$Specific)
buildings_test <- buildings_test[, !(colnames(buildings_test) %in% c("Specific","Generic"))]
To avoid having to push too much data into Git, we export dataset buildings_test and import it back using the code chunk below.
write_csv(buildings_test, "data/final_buildings.csv")
final_buildings <- read_csv("data/final_buildings.csv")
Data For Traffic To Each Building
Now we have to wrangle the data for participant activity in these buildings. We do so with the code chunk below.
# We convert the data into daily data. Handling hourly or even minute by minute data would be too computationally intense.
checkInJournal <- checkInJournal %>%
mutate(timestamp = as.Date(timestamp, format = "%m/%d/%y"))
checkInJournal$timestamp <- format(as.Date(checkInJournal$timestamp), "%Y-%m")
# Next we remove the columns participantId and venueType since we already have that data in the dataset final_buildings
checkInJournal <- checkInJournal[, !(colnames(checkInJournal) %in% c("participantId", "venueType"))]
# Now we get the monthly usage of each building
CheckInJournal2 <- checkInJournal %>%
group_by(timestamp, venueId) %>%
summarise(monthlyUsage = n())
# Now we create datasets for each month's data usage and rename the columns
Mar22 <- filter(CheckInJournal2, timestamp == "2022-03")
names(Mar22)[names(Mar22) == 'monthlyUsage'] <- "Mar22"
names(Mar22)[names(Mar22) == 'venueId'] <- "buildingId"
Mar22 <- Mar22[, !(colnames(Mar22) %in% c("timestamp"))]
Apr22 <- filter(CheckInJournal2, timestamp == "2022-04")
names(Apr22)[names(Apr22) == 'monthlyUsage'] <- "Apr22"
names(Apr22)[names(Apr22) == 'venueId'] <- "buildingId"
Apr22 <- Apr22[, !(colnames(Apr22) %in% c("timestamp"))]
May22 <- filter(CheckInJournal2, timestamp == "2022-05")
names(May22)[names(May22) == 'monthlyUsage'] <- "May22"
names(May22)[names(May22) == 'venueId'] <- "buildingId"
May22 <- May22[, !(colnames(May22) %in% c("timestamp"))]
Jun22 <- filter(CheckInJournal2, timestamp == "2022-06")
names(Jun22)[names(Jun22) == 'monthlyUsage'] <- "Jun22"
names(Jun22)[names(Jun22) == 'venueId'] <- "buildingId"
Jun22 <- Jun22[, !(colnames(Jun22) %in% c("timestamp"))]
Jul22 <- filter(CheckInJournal2, timestamp == "2022-07")
names(Jul22)[names(Jul22) == 'monthlyUsage'] <- "Jul22"
names(Jul22)[names(Jul22) == 'venueId'] <- "buildingId"
Jul22 <- Jul22[, !(colnames(Jul22) %in% c("timestamp"))]
Aug22 <- filter(CheckInJournal2, timestamp == "2022-08")
names(Aug22)[names(Aug22) == 'monthlyUsage'] <- "Aug22"
names(Aug22)[names(Aug22) == 'venueId'] <- "buildingId"
Aug22 <- Aug22[, !(colnames(Aug22) %in% c("timestamp"))]
Sep22 <- filter(CheckInJournal2, timestamp == "2022-09")
names(Sep22)[names(Sep22) == 'monthlyUsage'] <- "Sep22"
names(Sep22)[names(Sep22) == 'venueId'] <- "buildingId"
Sep22 <- Sep22[, !(colnames(Sep22) %in% c("timestamp"))]
Oct22 <- filter(CheckInJournal2, timestamp == "2022-10")
names(Oct22)[names(Oct22) == 'monthlyUsage'] <- "Oct22"
names(Oct22)[names(Oct22) == 'venueId'] <- "buildingId"
Oct22 <- Oct22[, !(colnames(Oct22) %in% c("timestamp"))]
Nov22 <- filter(CheckInJournal2, timestamp == "2022-11")
names(Nov22)[names(Nov22) == 'monthlyUsage'] <- "Nov22"
names(Nov22)[names(Nov22) == 'venueId'] <- "buildingId"
Nov22 <- Nov22[, !(colnames(Nov22) %in% c("timestamp"))]
Dec22 <- filter(CheckInJournal2, timestamp == "2022-12")
names(Dec22)[names(Dec22) == 'monthlyUsage'] <- "Dec22"
names(Dec22)[names(Dec22) == 'venueId'] <- "buildingId"
Dec22 <- Dec22[, !(colnames(Dec22) %in% c("timestamp"))]
Jan23 <- filter(CheckInJournal2, timestamp == "2023-01")
names(Jan23)[names(Jan23) == 'monthlyUsage'] <- "Jan23"
names(Jan23)[names(Jan23) == 'venueId'] <- "buildingId"
Jan23 <- Jan23[, !(colnames(Jan23) %in% c("timestamp"))]
Feb23 <- filter(CheckInJournal2, timestamp == "2023-02")
names(Feb23)[names(Feb23) == 'monthlyUsage'] <- "Feb23"
names(Feb23)[names(Feb23) == 'venueId'] <- "buildingId"
Feb23 <- Feb23[, !(colnames(Feb23) %in% c("timestamp"))]
Mar23 <- filter(CheckInJournal2, timestamp == "2023-03")
names(Mar23)[names(Mar23) == 'monthlyUsage'] <- "Mar23"
names(Mar23)[names(Mar23) == 'venueId'] <- "buildingId"
Mar23 <- Mar23[, !(colnames(Mar23) %in% c("timestamp"))]
Apr23 <- filter(CheckInJournal2, timestamp == "2023-04")
names(Apr23)[names(Apr23) == 'monthlyUsage'] <- "Apr23"
names(Apr23)[names(Apr23) == 'venueId'] <- "buildingId"
Apr23 <- Apr23[, !(colnames(Apr23) %in% c("timestamp"))]
May23 <- filter(CheckInJournal2, timestamp == "2023-05")
names(May23)[names(May23) == 'monthlyUsage'] <- "May23"
names(May23)[names(May23) == 'venueId'] <- "buildingId"
May23 <- May23[, !(colnames(May23) %in% c("timestamp"))]
Data For The Types of Buildings, Their Max Occupancy and Traffic To Each Building
Now we combine the datasets that we have obtained above in order to compare the columns maxOccupancy, with the monthly usage.
final_buildings <- merge(x = final_buildings, y = Mar22, by = "buildingId", all.x = TRUE)
final_buildings <- merge(x = final_buildings, y = Apr22, by = "buildingId", all.x = TRUE)
final_buildings <- merge(x = final_buildings, y = May22, by = "buildingId", all.x = TRUE)
final_buildings <- merge(x = final_buildings, y = Jun22, by = "buildingId", all.x = TRUE)
final_buildings <- merge(x = final_buildings, y = Jul22, by = "buildingId", all.x = TRUE)
final_buildings <- merge(x = final_buildings, y = Aug22, by = "buildingId", all.x = TRUE)
final_buildings <- merge(x = final_buildings, y = Sep22, by = "buildingId", all.x = TRUE)
final_buildings <- merge(x = final_buildings, y = Oct22, by = "buildingId", all.x = TRUE)
final_buildings <- merge(x = final_buildings, y = Nov22, by = "buildingId", all.x = TRUE)
final_buildings <- merge(x = final_buildings, y = Dec22, by = "buildingId", all.x = TRUE)
final_buildings <- merge(x = final_buildings, y = Jan23, by = "buildingId", all.x = TRUE)
final_buildings <- merge(x = final_buildings, y = Feb23, by = "buildingId", all.x = TRUE)
final_buildings <- merge(x = final_buildings, y = Mar23, by = "buildingId", all.x = TRUE)
final_buildings <- merge(x = final_buildings, y = Apr23, by = "buildingId", all.x = TRUE)
final_buildings <- merge(x = final_buildings, y = May23, by = "buildingId", all.x = TRUE)
And we bring the monthly usage down to daily usage to make the comparison a little easier.
final_buildings <- final_buildings %>%
mutate(Mar22Daily = Mar22 / 31) %>%
mutate(Apr22Daily = Apr22 / 30) %>%
mutate(May22Daily = May22 / 31) %>%
mutate(Jun22Daily = Jun22 / 30) %>%
mutate(Jul22Daily = Jul22 / 31) %>%
mutate(Aug22Daily = Aug22 / 31) %>%
mutate(Sep22Daily = Sep22 / 30) %>%
mutate(Oct22Daily = Oct22 / 31) %>%
mutate(Nov22Daily = Nov22 / 30) %>%
mutate(Dec22Daily = Dec22 / 31) %>%
mutate(Jan23Daily = Jan23 / 31) %>%
mutate(Feb23Daily = Feb23 / 28) %>%
mutate(Mar23Daily = Mar23 / 31) %>%
mutate(Apr23Daily = Apr23 / 30) %>%
mutate(May23Daily = May23 / 31)
Let us add one last column to show how close the building is to being overoccupied.
final_buildings <- final_buildings %>%
mutate(overOccupiedMar22Daily =
if_else(
(is.na(final_buildings$maxOccupancy) | is.na(final_buildings$Mar22Daily)),
0.00,
(final_buildings$Mar22Daily/ final_buildings$maxOccupancy))) %>%
mutate(overOccupiedApr22Daily =
if_else(
(is.na(final_buildings$maxOccupancy) | is.na(final_buildings$Apr22Daily)),
0.00,
(final_buildings$Apr22Daily/ final_buildings$maxOccupancy))) %>%
mutate(overOccupiedMay22Daily =
if_else(
(is.na(final_buildings$maxOccupancy) | is.na(final_buildings$May22Daily)),
0.00,
(final_buildings$May22Daily/ final_buildings$maxOccupancy))) %>%
mutate(overOccupiedJun22Daily =
if_else(
(is.na(final_buildings$maxOccupancy) | is.na(final_buildings$Jun22Daily)),
0.00,
(final_buildings$Jun22Daily/ final_buildings$maxOccupancy))) %>%
mutate(overOccupiedJul22Daily =
if_else(
(is.na(final_buildings$maxOccupancy) | is.na(final_buildings$Jul22Daily)),
0.00,
(final_buildings$Jul22Daily/ final_buildings$maxOccupancy))) %>%
mutate(overOccupiedAug22Daily =
if_else(
(is.na(final_buildings$maxOccupancy) | is.na(final_buildings$Aug22Daily)),
0.00,
(final_buildings$Aug22Daily/ final_buildings$maxOccupancy))) %>%
mutate(overOccupiedSep22Daily =
if_else(
(is.na(final_buildings$maxOccupancy) | is.na(final_buildings$Sep22Daily)),
0.00,
(final_buildings$Sep22Daily/ final_buildings$maxOccupancy))) %>%
mutate(overOccupiedOct22Daily =
if_else(
(is.na(final_buildings$maxOccupancy) | is.na(final_buildings$Oct22Daily)),
0.00,
(final_buildings$Oct22Daily/ final_buildings$maxOccupancy))) %>%
mutate(overOccupiedNov22Daily =
if_else(
(is.na(final_buildings$maxOccupancy) | is.na(final_buildings$Nov22Daily)),
0.00,
(final_buildings$Nov22Daily/ final_buildings$maxOccupancy))) %>%
mutate(overOccupiedDec22Daily =
if_else(
(is.na(final_buildings$maxOccupancy) | is.na(final_buildings$Dec22Daily)),
0.00,
(final_buildings$Dec22Daily/ final_buildings$maxOccupancy))) %>%
mutate(overOccupiedJan23Daily =
if_else(
(is.na(final_buildings$maxOccupancy) | is.na(final_buildings$Jan23Daily)),
0.00,
(final_buildings$Jan23Daily/ final_buildings$maxOccupancy))) %>%
mutate(overOccupiedFeb23Daily =
if_else(
(is.na(final_buildings$maxOccupancy) | is.na(final_buildings$Feb23Daily)),
0.00,
(final_buildings$Feb23Daily/ final_buildings$maxOccupancy))) %>%
mutate(overOccupiedApr23Daily =
if_else(
(is.na(final_buildings$maxOccupancy) | is.na(final_buildings$Apr23Daily)),
0.00,
(final_buildings$Apr23Daily/ final_buildings$maxOccupancy))) %>%
mutate(overOccupiedMay23Daily =
if_else(
(is.na(final_buildings$maxOccupancy) | is.na(final_buildings$May23Daily)),
0.00,
(final_buildings$May23Daily/ final_buildings$maxOccupancy)))
# Let us make sure there are no duplicates
final_buildings <- unique(final_buildings)
Let us export the dataset final_buildings and import it again to prevent git from crashing.
write_csv(final_buildings, "data/final_buildings_2.csv")
Please note that some changes were made to the CSV final_buildings_2. Hence, there is a change in the versioning. However, it was too minor to add here.
final_final_buildings <- read_sf("data/final_buildings_4.csv",
options = "GEOM_POSSIBLE_NAMES=location")
Visualisation
Answering Question 1
The code chunk below plots the building polygon features by using tm_polygon().
tmap_mode("plot")
tm_shape(final_final_buildings)+
tm_polygons(col = "buildingType",
palette="Accent",
border.col = "black",
border.alpha = .5,
border.lwd = 0.5)
tmap_mode("plot")
Now let us get use the same colour palette to identify the buildings, while using the size of circles to represent how social the areas are each month.
March 2022
final_final_buildings$A = as.numeric(final_final_buildings$A)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "A",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
Apr 2022
final_final_buildings$B = as.numeric(final_final_buildings$B)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "B",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
May 2022
final_final_buildings$C = as.numeric(final_final_buildings$C)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "C",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
June 2022
final_final_buildings$D = as.numeric(final_final_buildings$D)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "D",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
July 2022
final_final_buildings$E = as.numeric(final_final_buildings$E)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "E",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
August 2022
final_final_buildings$F = as.numeric(final_final_buildings$F)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "F",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
September 2022
final_final_buildings$G = as.numeric(final_final_buildings$G)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "G",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
October 2022
final_final_buildings$H = as.numeric(final_final_buildings$H)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "H",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
November 2022
final_final_buildings$I = as.numeric(final_final_buildings$I)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "I",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
December 2022
final_final_buildings$J = as.numeric(final_final_buildings$J)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "J",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
January 2023
final_final_buildings$K = as.numeric(final_final_buildings$K)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "K",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
February 2023
final_final_buildings$L = as.numeric(final_final_buildings$L)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "L",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
March 2023
final_final_buildings$M = as.numeric(final_final_buildings$M)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "M",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
April 2023
final_final_buildings$N = as.numeric(final_final_buildings$N)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "N",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
May 2023
final_final_buildings$O = as.numeric(final_final_buildings$O)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "O",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
Insight
To answer the first question:
Based on the tmaps over time, it seems that the social areas are quite consistent. If we base it off the total number of significantly large bubbles, the most social areas in descending order are:
- Apartments
- Employer
- Commercial
Answering Question 2
We essentially do the same thing as Question 1 but using the over-occupancy metrics such as overOccupiedMar22Daily.
March 2022
final_final_buildings$overOccupiedMar22Daily = as.numeric(final_final_buildings$overOccupiedMar22Daily)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "overOccupiedMar22Daily",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
Apr 2022
final_final_buildings$overOccupiedApr22Daily = as.numeric(final_final_buildings$overOccupiedApr22Daily)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "overOccupiedApr22Daily",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
May 2022
final_final_buildings$overOccupiedMay22Daily = as.numeric(final_final_buildings$overOccupiedMay22Daily)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "overOccupiedMay22Daily",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
June 2022
final_final_buildings$overOccupiedJun22Daily = as.numeric(final_final_buildings$overOccupiedJun22Daily)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "overOccupiedJun22Daily",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
July 2022
final_final_buildings$overOccupiedJul22Daily = as.numeric(final_final_buildings$overOccupiedJul22Daily)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "overOccupiedJul22Daily",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
August 2022
final_final_buildings$overOccupiedAug22Daily = as.numeric(final_final_buildings$overOccupiedAug22Daily)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "overOccupiedAug22Daily",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
September 2022
final_final_buildings$overOccupiedSep22Daily = as.numeric(final_final_buildings$overOccupiedSep22Daily)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "overOccupiedSep22Daily",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
October 2022
final_final_buildings$overOccupiedOct22Daily = as.numeric(final_final_buildings$overOccupiedOct22Daily)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "overOccupiedOct22Daily",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
November 2022
final_final_buildings$overOccupiedNov22Daily = as.numeric(final_final_buildings$overOccupiedNov22Daily)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "overOccupiedNov22Daily",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
December 2022
final_final_buildings$overOccupiedDec22Daily = as.numeric(final_final_buildings$overOccupiedDec22Daily)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "overOccupiedDec22Daily",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
January 2023
final_final_buildings$overOccupiedJan23Daily = as.numeric(final_final_buildings$overOccupiedJan23Daily)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "overOccupiedJan23Daily",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
February 2023
final_final_buildings$overOccupiedFeb23Daily = as.numeric(final_final_buildings$overOccupiedFeb23Daily)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "overOccupiedFeb23Daily",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
April 2023
final_final_buildings$overOccupiedApr23Daily = as.numeric(final_final_buildings$overOccupiedApr23Daily)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "overOccupiedApr23Daily",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
May 2023
final_final_buildings$overOccupiedMay23Daily = as.numeric(final_final_buildings$overOccupiedMay23Daily)
tm_shape(final_final_buildings)+
tm_bubbles(col = "buildingType",
alpha = 0.8,
n = 6,
style = "jenks",
palette="Accent",
size = "overOccupiedMay23Daily",
scale = 0.8,
border.col = "black",
border.lwd = 0.5)
Insight
To answer the second question:
Based on the tmaps over time, it does seem like that there are minimal bottlenecks. And that the worst ones are experienced by apartments since the size of the “over-occupancy” bubbles are the highest for them.
Review
I believe that the visualisation hardly revealed any information because the difference in magnitude between the most “social” and most severely “bottleneck” areas, and the other areas was too drastic.
To solve this problem, I could have implemented an additional layer of visual analysis by capping the upper limit for the level of “social” or “bottleneck” in buildings in Engagement. Instead of producing 24 different graphs, this question was also a great opportunity for dynamic analysis.